9 個模型已上線

AI 圖片生成器

多個前沿模型匯於一個生成器——Nano Banana 2、Nano Banana 2 Lite、Nano Banana Pro、GPT Image 2、FLUX.2 Dev/Pro/Max，以及 Seedream 5.0 Pro/Lite。把文字提示與參考圖，變成從 1K 草稿到 4K 輸出的視覺成品。

一個生成器，匯聚領先圖片模型

為每張視覺作品挑對引擎——畫質、速度，或風格控制。

GPT Image 2

清晰文字與精確尺寸——從極簡參數到精細輸出

1K2K4K

Nano Banana 2 Lite

家族最快檔——1K 快速出圖，成本最低

1:11:41:82:3+11

Nano Banana 2

超寬長寬比，最高 4K 細節

1K2K4K

Nano Banana Pro

專業級細節與多參考圖引導，最高 4K

1K2K4K

FLUX.2 Dev

FLUX.2 快速草稿，完整保留提示詞細節

auto1:116:93:2+7

FLUX.2 Pro

FLUX.2 高品質輸出，支援解析度檔位

0.5MP1MP2MP4MP

FLUX.2 Max

FLUX.2 頂級畫質，滿足最挑剔的渲染需求

0.5MP1MP2MP4MP

Seedream 5.0 Lite

新上線

ByteDance Seedream 輕量版——低成本產出 2K–4K 高解析圖

2K3K4K

Seedream 5.0 Pro

新上線

ByteDance 旗艦——每次一張高品質 2K 圖片

1K2K

FLUX 3

即將上線

BFL 多模態模型——複雜提示詞與多語言文字渲染

1:12:33:23:4+6
了解更多

Muse Image

即將上線

Meta 的智能體圖片模型——開放當天啟用

1:12:33:23:4+6
了解更多

Grok Imagine 2.0

即將上線

xAI 旗艦——Aurora-2 引擎、4MP 原生輸出、Quality 模式

1:12:33:23:4+6
了解更多

已上線圖片模型比較

所有模型共用同一個生成器、同一份積分與同一套工作流程——真正的差異在解析度、長寬比與能帶進去的參考圖數量。下方規格直接讀取自生成器設定。

模型最高解析度長寬比參考圖單次出圖數適合場景
GPT Image 21K / 2K / 4K16 種預設最多 16最多 10清晰文字、單次最多 10 張
Nano Banana 2 Lite1K15 種預設最多 101快速草稿與社群大量出圖
Nano Banana 21K / 2K / 4K15 種預設最多 141超寬全景與細節豐富的 4K 場景
Nano Banana Pro1K / 2K / 4K11 種預設最多 81多參考圖引導的商品攝影
FLUX.2 Dev最長邊 1440px（預設或自訂）11 種預設最多 51FLUX.2 快速草稿與提示詞探索
FLUX.2 Pro0.5–4 MP（預設或自訂最高 2048px）11 種預設最多 81解析度可控的高品質 FLUX.2 出圖
FLUX.2 Max0.5–4 MP（預設或自訂最高 2048px）11 種預設最多 81複雜場景下的 FLUX.2 頂級畫質
Seedream 5.0 Lite2K / 3K / 4K8 種預設最多 101低成本 2K–4K Seedream 出圖
Seedream 5.0 Pro1K / 1.5K / 2K8 種預設最多 101單次一張高品質 Seedream 輸出

以上全部都在上方的生成器裡跑——在模型下拉選單挑一個，參數面板就會自動對應調整。

靈感畫廊

都是在 Molyin 上生成的真實圖片——挑一張，改成你自己的。

1.A cute, stylized 3D digital character of a young woman with soft fair skin and rosy cheeks, winking playfully at the camera. She has large expressive amber eyes with a golden ring, long eyelashes, and a small gentle smile. Her curly strawberry-blonde hair is styled in a fluffy high bun with loose strands, wearing white over-ear headphones and pastel pink sunglasses resting on her head. She is wearing a cozy oversized pink knitted cardigan, a fluffy cream crop top, and light beige lounge shorts with a small pink bow. The lighting is warm, soft and cinematic with a pastel bokeh background, dreamy atmosphere, shallow depth of field, ultra-detailed textures, soft global illumination, smooth skin shading, Pixar / Disney-style 3D render, ultra-high resolution, 4K, sharp focus, clean, vibrant, aesthetic, studio quality. 2. A cute stylized 3D cartoon girl character, waist-up portrait, soft fair skin with rosy cheeks, big expressive brown eyes, wearing round pink translucent glasses, smiling gently and looking slightly to the side. Curly light-brown hair tied in a fluffy messy bun with loose strands framing her face. Wearing large white over-ear headphones. Outfit: soft fuzzy pastel pink zip-up hoodie worn open, white sporty crop top underneath, peach-pink lounge pants. Slim youthful proportions, smooth soft shading. Warm pastel rainbow gradient background with soft bokeh glow. Dreamy lighting, soft rim light, shallow depth of field. Ultra detailed textures, high quality Pixar / Disney style 3D render, cinematic lighting, ultra HD, 4K, sharp focus, clean, aesthetic.
  • Nano Banana 2
  • Text to image
  • 3:4
Create an ultra-premium advertising poster for Minute Maid, centered on a hyper-real grape-flavor Minute Maid juice package that appears physically fused with the trunk of a living grapevine tree, as if the product itself has grown directly from the plant. The image must feel fresh, natural, luminous, and commercially powerful, with a deep organic integration between the package and the tree structure. The final result should look like a flagship Minute Maid campaign where the drink is not placed in nature, but born from it. Core concept: The hero is a hyper-real grape-flavor Minute Maid carton or bottle standing upright in the center of a lush natural vineyard-garden environment. The key visual idea is that the packaging is not separate from the tree. It must look like it is part of the trunk itself, emerging from bark, roots, and vine wood as a living extension of the grape plant. The lower and side edges of the package should merge naturally into the trunk base, root system, and vine-wrapped bark, creating the impression that the product is a cultivated fruit-bearing life form. The message must feel immediate and visually unforgettable: Minute Maid grape juice is grown directly from living nature. Composition: Use a vertical premium poster layout with the Minute Maid product placed prominently in the lower-middle center, integrated into the core of a vine-covered tree or thick grapevine trunk. The tree must rise behind and around the package, with the bark, roots, and woody vine structure wrapping the product naturally. In the foreground, use grass, vineyard soil, small flowers, tendrils, and root extensions to anchor the fusion. Above and around the product, grapevine branches should carry clusters of ripe grapes and broad leaves, creating a natural canopy and flavor crown. The background should open into a soft vineyard or garden landscape with subtle depth and cinematic air. Product realism: Render the Minute Maid grape-flavor packaging with absolute realism and premium fidelity: accurate carton or bottle structure, clear Minute Maid branding, precise cap details, clean printed grape flavor identity, subtle reflections, realistic material texture, and strong commercial sharpness. The package must remain the most polished and readable object in the image, even while fused with the organic tree form. Tree-package fusion: The fusion between packaging and tree must be strong and unmistakable. The package should appear embedded in or emerging from the trunk, with bark textures, root structures, vine wood, and grape tendrils partially wrapping and blending around the lower edges and side planes. The transition must feel elegant and believable, not grotesque or random. The viewer should immediately understand that the product is part of the tree’s living structure. Grape flavor fidelity: Use only grapevine logic for the botanical system: abundant grape clusters, rich purple and deep red grapes, healthy vine leaves, tendrils, woody vine branches, and vineyard-like natural richness. The grapes must feel ripe, juicy, sunlit, and luxurious. All visible fruit must clearly reinforce grape flavor identity. Environment: Create a clean cinematic vineyard-garden environment: lush grass, soft soil, tiny flowers, vine roots, warm natural air, distant vineyard greenery, and subtle morning or late-afternoon light haze. The setting should feel believable but idealized, clean yet abundant, premium rather than rustic-messy. Background: Use a luminous natural vineyard backdrop with soft sky opening, distant vine rows or orchard-like depth, warm haze, and elegant atmospheric layering. The background must remain soft enough to keep the package-tree fusion dominant. Typography: Use an elegant minimal English-led typography system. Suggested text direction: Main English title: “GROWN INTO FLAVOR” or “ROOTED IN GRAPE” Supporting English copy: “From the vine. From the trunk. Into every sip.” Typography should be minimal, clean, and premium, pl
  • gpt-image-2-basic
  • Text to image
  • 3:4
Create an ultra-vibrant commercial poster for M&M’s chocolate candies, featuring a highly realistic mini M&M’s tube package at the exact center as the absolute hero object. The image should combine one realistic product with an explosive hand-drawn illustrated background, creating a youthful, saturated, energetic, and extremely lively candy campaign poster. Main composition: Place one realistic mini M&M’s tube package exactly at the center of the image. The package should be a slim portable candy tube, compact and upright, with a rounded cap and a clean cylindrical body, clearly designed as a mini snack tube for M&M’s. It must feel highly realistic, glossy, tactile, and instantly recognizable, with vivid M&M’s logo, colorful branding, premium printed label wrap, sharp reflections, and strong commercial product rendering. The mini tube should be the only fully realistic object in the composition and must dominate the center. Packaging direction: The tube should feel playful, portable, collectible, and snackable, like a convenient mini candy tube sold at retail checkout counters. The cap can be bold red, yellow, blue, or white. The body should clearly show M&M’s identity and may include a transparent section revealing colorful candies inside, or a bright fully printed wrapper with bold brand graphics. The tube should feel modern, cute, and highly display-friendly. Background concept: Behind the mini tube, create an explosive illustrated universe that radiates outward like a burst of candy energy. The background should not be realistic. Instead, it should be a dense, dynamic, hand-drawn graphic world made of bold black outlines, candy splashes, bounce arcs, swirls, pop-art shapes, looping doodles, comic impact bursts, and rhythmic motion graphics. The illustration should erupt from behind the tube and wrap around it in a circular energetic explosion. Illustrated elements: Build the illustrated background from M&M’s-inspired candy energy: flying colorful chocolate beans in red, yellow, blue, green, orange, and brown, candy trails, sugar splash shapes, confetti, comic impact bursts, lightning bolts, stars, hearts, bounce curves, looping lines, speech-bubble icons, playful smiley doodles, cheerful motion marks, curved wave forms, mini balloons, striped speed lines, roller-skate energy, fun geometric symbols. The illustrations should feel premium, bold, rhythmic, and highly art-directed, not messy childish scribbles. Style contrast: The mini tube must feel solid, glossy, realistic, and sharply rendered, while the illustrated universe behind it feels wild, colorful, playful, and explosive. This contrast between realism and hand-drawn candy chaos is the key visual hook. The product remains the visual anchor at all times. Color palette: Use a highly saturated M&M’s palette: bright candy red, electric blue, sunny yellow, vivid green, orange, chocolate brown, clean white highlights, and bold black line work. Choose one strong dominant background field color such as vivid sky blue, rich yellow, or bright red, then let the multicolor candy illustration explode around the central mini tube. Motion and energy: The poster must feel like candy energy is bursting outward from the mini tube with bounce, swirl, pop, splash, spin, and playful impact. The overall mood should be dynamic, joyful, fast, pop, and impossible to ignore. Typography: Keep typography concise and energetic, placed cleanly without cluttering the composition. Use playful premium lettering, rounded and graphic, not ugly heavy black fonts. Include text such as: “M&M’s” “MINI TUBE, MAX FUN” “POP THE COLOR” “SHAKE. SNAP. PLAY.” Optional artistic Chinese text: “M&M’s 玛氏巧克力豆” “迷你小管装” “小小一管，快乐爆开” “随手一拿，彩色开玩” Keep text short, bold, youthful, and balanced with the central tube and illustrated explosion. Mood: playful, saturated, dynamic, youthful, colorful, pop-art, joyful, energetic, collectible, premium, unforgettable. Rendering: hyper-detailed realistic mini candy tube p
  • gpt-image-2-basic
  • Text to image
  • 3:4
Create a vintage-inspired travel poster in a layered 3D paper-cut collage style, designed like a worn traveler’s scrapbook page. Use torn parchment paper with sepia textures, realistic paper layers, soft drop shadows, and an antique map subtly visible beneath. Arrange iconic landmarks in a balanced composition with delicate handwritten labels and arrows, complemented by lush greenery, cobblestone streets, and charming local architecture. Add whimsical travel details such as a paper airplane with a dotted flight path, flying birds, vintage postage stamps, postal marks, and a compass rose. Finish with elegant serif typography featuring the destination name in large spaced letters and a smaller line listing notable cities below. The overall mood should feel nostalgic, handcrafted, richly detailed, timeless, and suitable for a premium vintage travel poster.
  • gpt-image-2-basic
  • Text to image
  • 4:3
Universal Plug-in Prompt A full-body high-fashion editorial studio photograph of a female model wearing an avant-garde couture gown constructed entirely from [PRIMARY MATERIAL]. The gown features [SILHOUETTE], enhanced with [STRUCTURAL DESIGN], showcasing intricate craftsmanship and dramatic sculptural volume. Behind the model stands an oversized [BACKGROUND INSTALLATION], designed using [BACKGROUND MATERIAL], visually echoing the gown while creating a cohesive artistic composition. The color palette transitions from [COLOR A] through [COLOR B] into [COLOR C], maintaining smooth gradients and harmonious visual balance. The model is captured from [CAMERA ANGLE] in [POSE], with [HAIRSTYLE], expressing elegance, confidence, and refined artistic presence. Professional luxury fashion photography. Museum-quality couture. Sharp editorial styling. Exceptional material realism. Highly detailed textures. Crisp edges. Dramatic dimensional depth. Sculptural lighting tailored to emphasize every fold, bead, petal, pleat, or layered surface. Minimalist premium studio environment featuring [BACKGROUND COLOR], [FLOOR TYPE], and cinematic composition. Balanced negative space. High contrast with soft shadow transitions. Ultra-clean luxury aesthetic. Cheat Sheet [PRIMARY MATERIAL] → Beads, pom-poms, paper flowers, feathers, shells, glass, ribbons [SILHOUETTE] → Ball gown, trumpet, cape, circular skirt, cathedral train [STRUCTURAL DESIGN] → Pleats, layered ruffles, radial folds, halos, geometric panels [BACKGROUND INSTALLATION] → Giant fan, flower wall, halo, sun, wings, sculpture [BACKGROUND MATERIAL] → Paper flowers, beads, fabric folds, petals, metallic elements [COLOR A/B/C] → Main gradient or palette [CAMERA ANGLE] → Front, side, back, three-quarter, profile [POSE] → Arms extended, holding fan, walking, profile stance, graceful turn [HAIRSTYLE] → Bun, low bun, ponytail, loose waves, braided updo [BACKGROUND COLOR] → Charcoal, ivory, grey, black, cream [FLOOR TYPE] → Studio floor, wood planks, textured concrete, seamless backdrop
  • gpt-image-2-basic
  • Text to image
  • 3:4
Clean, modern typographic travel poster in a 3:2 landscape format. The words "NEW YORK" — seven large, bold, uppercase geometric sans-serif letters set on a single line (or "NEW" stacked above "YORK" if better balanced) — span the full width and ARE the entire composition. Every illustrated element lives strictly INSIDE the letterforms, as if each letter is a window cut out of white paper revealing a scene; nothing spills outside the letter silhouettes. Outside the letters: pure, empty, flat white background — no objects, no cast shadows, no decoration. — WHAT FILLS EACH LETTER (NYC-accurate, clipped to the letter shape) — N : the Statue of Liberty — standing torch raised, following one vertical of the N. E : stacked Manhattan brownstone facades with fire escapes nested in the three horizontal bars. W : the Brooklyn Bridge — its two Gothic stone arches and suspension cables spanning the W's zig-zag, a small boat on the East River below. Y : the Empire State Building — the Art-Deco spire cradled in the fork of the Y. O : the ring frames Times Square — glowing billboards, a yellow taxi cab, a scramble of tiny pedestrians. R : the Flatiron Building's distinctive triangular wedge silhouette inside the R. K : Central Park greenery and the One World Trade Center spire slicing along the K's diagonals; a subway car along the base. Small city details (a yellow cab, a hot-dog cart with faint steam, a pigeon, a subway roundel, drifting clouds) appear ONLY where they fit within a letter's interior. Keep full legibility of "NEW YORK". — STYLE — Elegant flat vector illustration, crisp geometric shapes, minimal detail, clean outlines, subtle shadows contained inside the letters only. Limited palette: deep navy, warm cream, muted red, soft gray-blue — a restrained warm-yellow accent used ONLY on the taxis / Times Square glow. Optional minimal tagline beneath in wide-spaced small capitals: "THE CITY THAT NEVER SLEEPS", thin and understated. Perfectly balanced centered composition. premium flat vector, minimalist travel poster, geometric illustration, editorial tourism branding, clean typography, high contrast, ultra-sharp lines, museum-quality print, scalable SVG aesthetic, 8K.
  • gpt-image-2-basic
  • Text to image
  • 4:3

為圖片創作者而生

從提示詞到成品圖片，你要的一切都在這裡。

文生圖

輸入提示詞，幾秒內拿到成品。Nano Banana 2 與 GPT Image 2 聽得懂複雜指令——主體、風格、光線，甚至圖內清晰可讀的文字。

圖生圖

上傳參考圖——GPT Image 2 最多 16 張——融合成一張結果：幫產品換風格、混搭角色，或讓同一張臉貫穿整個系列。

最高 4K 解析度

先用 1K 打草稿，再用同一條提示詞以 2K 或 4K 重跑，直接滿足印刷、hero 橫幅與大幅輸出——不用另外跑放大。

長寬比全覆蓋

從 1:1 貼文到 21:9 電影畫幅，再到極限 8:1 全景——選預設，或自動對齊參考圖的長寬比。

無浮水印，可商用

付費方案下載即為無浮水印的乾淨檔案，商用授權涵蓋客戶專案、廣告與商品頁。

一份積分全站通用

積分在圖片與影片生成間通用，送出前就能看到確切消耗，生成失敗自動退還積分。

使用方法

從提示詞到 4K 下載，只要三步。

描述你的構想

輸入文字提示詞或上傳參考圖，再選模型——參數面板會自動對應它的能力。

生成圖片

選好長寬比與解析度，確認積分預估後送出。結果幾秒內就出現在生成器下方。

下載與使用

下載圖片、用圖生圖繼續改，或用同一條提示詞以更高解析度重跑。

AI 圖片生成器能做什麼？

創作者每天在 Molyin 上跑的六種工作流程——每一種都從一句白話提示詞開始。

社群貼文與封面圖

為 Instagram、TikTok 與 YouTube 封面做出讓人停下手指的視覺——1:1、4:5、9:16 預設一步到位。

AI 商品攝影

一張商品參考圖，變出棚拍級目錄照、生活情境圖與季節活動素材——不必進棚。

廣告素材與橫幅

用 AI 的速度試遍鉤子、風格與版型——再把勝出的版本以 4K 匯出，用於展示廣告、印刷或 21:9 超寬版位。

簡報與說明圖

把抽象概念變成清楚易懂的插圖，服務課程、講義與簡報。

品牌與網頁設計

與你的視覺系統一致的 hero 圖、背景、圖示與樣機——需要時還能輸出乾淨背景。

動漫與遊戲概念圖

角色、場景與主視覺，服務世界觀建構、二創與預告片——整套素材風格統一。

什麼是 AI 圖片生成器？

AI 圖片生成器把一段文字描述或一張參考圖，在幾秒內變成一幅成品——不用相機、不用畫布、不用設計軟體。Nano Banana 2、GPT Image 2 這類模型在數十億組圖文配對上訓練，所以能把關於主體、風格、光線與構圖的自然語言，轉成新圖片裡的每一個像素。

Molyin 把這些前沿模型收進同一個提示框。你不必在多個工具之間來回切換帳號，只要寫一次想法，再挑最合適的引擎——最快的、最省的，還是細節最強的——然後在同一份積分下並排比較結果。兩種輸入模式涵蓋所有起點：

文生圖

從純文字出發。描述主體、風格、光線與構圖——例如 "a cyberpunk street at night, neon reflections on wet asphalt, cinematic"——模型就會建出整幅圖片。文生圖是探索概念、在沒有現成照片時產出素材、以及為一個點子一次生成多種變體的最快路徑。

圖生圖

把現有圖片當參考帶進來。上傳商品照、草圖或幾張風格幀，模型在保留關鍵要素的同時重新詮釋——非常適合幫產品換新裝、讓角色在整套素材中保持一致，或把一批視覺統一到同一種調性。

怎麼寫出更好的 AI 圖片提示詞

提示詞品質比其他任何設定都更能決定出圖品質。四個能穩定提升結果的習慣：

  • 先寫主體與場景 — 把主體放在第一個分句，再交代環境——"a ceramic espresso cup on a marble counter, morning light" 遠勝一堆沒有錨點的形容詞。
  • 點名媒材與風格 — "studio photograph"、"watercolor"、"anime key visual"、"3D render"——一個明確的風格詞，比十個模糊的詞更能定調整幅畫面。
  • 指揮鏡頭與光線 — "close-up"、"wide shot"、"golden hour"、"soft diffused light"、"dramatic backlight" 這類詞，正是模型在訓練時學到的攝影語言。
  • 1K 疊代，4K 收尾 — 先用 Nano Banana 2 Lite 或 FLUX.2 Dev 這類快模型探索想法，再把勝出的提示詞放到 Nano Banana 2、GPT Image 2 或 FLUX.2 Max 上以 2K/4K 重跑，完成最終輸出。

該選哪個 AI 圖片模型？

Nano Banana 2 Lite 是速度擔當——最低積分成本的 1K 快速草稿，適合社群大量出圖與提示詞探索。Nano Banana 2 升級到 4K，長寬比一路拉到 8:1 全景。Nano Banana Pro 加入多參考圖引導，適合必須保持一致的產品與品牌工作。GPT Image 2 是圖裡需要清晰文字時的首選——同一個模型、兩種參數檔：Easy 檔用最簡單的設定涵蓋日常出圖，Advanced 檔解鎖精確像素尺寸與單次最多 10 張。FLUX.2 Dev 是體驗 FLUX.2 家族的最快入口，Pro 加入最高 4 MP 的顯式解析度控制，Max 則是複雜、高細節場景下的畫質天花板。Seedream 5.0 Pro 每次輸出一張高品質 2K 圖片，Seedream 5.0 Lite 則用更低積分成本輸出 2K–4K 的同款 Seedream 風格。

你不必先選邊站：所有模型都在同一個下拉選單裡，參數面板會隨你選的模型自動對應，送出前積分預估即時更新。可以看上方的規格比較表，或乾脆把同一條提示詞丟給兩個模型，用眼睛當裁判。

每張圖片都會存進你的雲端作品庫，付費方案解鎖無浮水印下載與商用授權，同一份積分也能驅動我們的 AI 影片生成器。到 定價頁比較方案——或回到上方，用註冊送的免費積分生成你的第一張圖片。

深入創作之道

來自創作現場的實戰指南，與每一幀背後的詞彙——提示詞、模型，以及解鎖它們的關鍵概念。

常見問題

關於在 Molyin 上生成 AI 圖片，你想知道的都在這裡。

現在就生成你的第一張 AI 圖片

註冊即送免費積分——多個前沿模型，一個提示框。

開始創作查看定價